REVIEW OF: Nuñez NA et al, J Affect Disord 2022;302:385–400

STUDY TYPE: Meta-analysis of ­placebo-controlled trials

Recent analyses have clarified an old debate. Should you switch antidepressants or augment when depression does not remit on an antidepressant? Augmentation is generally the more effective route, but which augmentation strategy is best? This network meta-analysis aimed to clarify that question.

The authors included 69 randomized controlled trials (RCTs) comprising patients with major depressive disorder ages 18–65 years, refractory to one or more antidepressants at an adequate dose. The studies compared an active drug with another drug or with placebo for up to 24 weeks. The authors excluded studies focusing on specific patient populations like those with psychotic features or with depressions that occurred in bipolar disorder, in the prenatal or postpartum periods, as part of a serious medical illness, or in the elderly. The primary outcome was response rate defined as a ≥50% decrease on a validated scale for depression (Hamilton Depression Rating Scale or Montgomery-Åsberg Depression Rating Scale). Secondary outcomes were remission and all-cause discontinuation rates. Results were expressed in relative risk (RR) with values >1 indicating superiority. Since RRs are approximate, they are listed below along with their confidence interval (CI), which represents the range of values that they are likely to fall into.

The augmentation strategies that stood out as significantly more effective than placebo were (in alphabetical order): aripiprazole (RR 1.57, CI 1.36–1.82); brexpiprazole (1.56, 1.15–2.11); cariprazine (1.20, 1.01–1.42); lisdexamfetamine (1.18, 1.03–1.37); lithium (1.25, 1.00–1.56); modafinil (1.26, 1.07–1.48); nortriptyline (2.05, 1.02–4.11); olanzapine (1.23, 1.00–1.50); quetiapine (1.34, 1.14–1.56); and T3 (1.90, 1.16–3.11). Those that did not reach significance were bupropion, buspirone, lamotrigine, methylphenidate, mirtazapine, pramipexole, risperidone, T4, and ziprasidone.

Any differences in effect were largely washed out by the CIs—particularly nortriptyline’s, which was based on only 23 patients. Among the options, the antipsychotics had the most studies and the largest sample sizes (aripiprazole, n = 1147; brexpiprazole, n = 599; cariprazine, n = 963; lithium, n = 469; modafinil, n = 284; nortriptyline, n = 23; T3, n = 114). Discontinuation rates were higher with cariprazine (1.72, 1.09–2.73), as well as with two drugs that failed the significance test—ziprasidone (20.12, 1.17–344.58) and mirtazapine (4.12, 1.97–8.63).

Network meta-analyses compare medication and placebo arms from different studies. This has inherent flaws, as placebo response varies with study design, even though measures of heterogeneity and inconsistency were nonsignificant in this analysis. The authors raised concern about one result—lisdexamfetamine—as earlier meta-analyses were not as favorable toward this stimulant. Indeed, the two large trials of lisdexamfetamine were negative while the two smaller ones were positive. The analysis also focused on pharmacotherapy, which leaves out other effective augmentation strategies like complementary-alternative (eg, SAMe, L-methylfolate), lifestyle (eg, aerobic exercise), psychotherapy, transcranial magnetic stimulation, ECT, and the ketamines.

CARLAT TAKE

This list is informative but not definitive, as there are many ways to slice the data. The antipsychotics are the best studied of the lot, but no strategy stands out as more effective than another. The choice, then, is best guided by balancing the risks of the medication with the severity of the patient’s depression.